CogniGraph / Docs

Find your
next answer.

Start with the question you are trying to answer. These CogniGraph guides explain the database, query language and integration model, then point to the engineering reference for exact contracts and operational details.

01 / Concepts & language

Understand what connects.

02 / Put it to work

Move from query to application.

The HTTP API is the integration boundary. CGQL accepts query text and bind values; dedicated endpoints cover full-text, vector, semantic, hybrid and graph-augmented retrieval. Lua combines database operations with server-side control flow and explicit execution limits.

When you need the exact contract

The running server exposes /openapi.yaml for its edition. The repository contains the CGQL grammar, validation restrictions and HTTP payload documentation. Use these when implementing or reviewing an integration.

03 / Beyond the first request

Know the operating boundaries.

Before sharing a deployment, work through authentication, configuration, backup and recovery. Both editions use Native storage, with resident and paged modes for different memory requirements. Each store has one owning process.

For evidence-bound construction or more than one tenant, start with the Enterprise overview. Its governance workflows have their own policy, review, artifact and deployment contracts. Plan around a single-writer deployment; replication is roadmap.

For a review of how changes and measured results are recorded, see engineering evidence.